Abstract
We compiled 4-day dietary records with digital photographic images of meals to investigate the variation of frequently consumed foods which could enable the development of a dish-based database for dietary assessment in epidemiological studies. We also evaluated the validity of using a digital photographic image as a tool to assess the dietary intake of a food group. Participants in the study were 144 men and women living in urban areas of Japan who were asked to take pictures with a digital camera of all foods and beverages they consumed during 4 consecutive days. Simultaneous dietary records were used as the reference. The dietary intake was independently calculated from the two methods. We found that the variation in intake was high for such seasoning as salt which indicated that the degree of measurement error may be high when a representative portion size for such seasoning was used in a dish-based database. The estimated intake level by two methods was similar for green-yellow vegetables, fruit, fish, meat, eggs, milk and other dairy products, whereas the over/under-estimation for other food groups was too great to be used for accurate assessment. In contrast, the validity for ranking individuals by the estimated multiple-day intake was relatively high, with the exception of beverages, seasonings and spices.
